SoRec: Social Recommendation Using Probabilistic Matrix Factorization

1
SoRec Social Recommendation UsingProbabilistic
Matrix Factorization

• Hao Ma
• Dept. of Computer Science Engineering
• The Chinese University of Hong Kong
• Co-work with Haixuan Yang, Michael R. Lyu and
  Irwin King

2
Background

• Do you have this experience?

3
(No Transcript)
4
Background

• Recommender Systems become more and more
  important

The number of Internet websites each year since
the Web's founding. From http//www.useit.com/aler
tbox/web-growth.html
5
Challenges

• Data sparsity problem

6
Number of Ratings per User
Extracted From Epinions.com 114,222 users,
754,987 items and 13,385,713 ratings
7
Challenges

• Traditional recommender systems ignore the social
  connections between users

Recommendations from friends
8
Challenges

• Yes, there is a correlation - from social
  networks to personal behavior on the web
• Parag Singla and Matthew Richardson (WWW08)
• Analyze the who talks to whom social network over
  10 million people with their related search
  results
• People who chat with each other are more likely
  to share the same or similar interests

9
Motivation

• To improve the recommendation accuracy and solve
  the data sparsity problem, users social network
  should be taken into consideration

10
Problem Definition
11
Social Network Graph Matrix Factorization
12
User-Item Rating Matrix Factorization
13
Social Recommendation
14
Gradient Descent
15
Complexity Analysis

• For the Objective Function
• For , the complexity is
• For , the complexity is
• For , the complexity is
• In general, the complexity of our method is
  linear with the observations in these two matrices

16
Related Work

• Combining content and link for classification
  using matrix factorization
• Shenghuo Zhu, et al. (SIGIR 2007)
• Differences
• Our method can deal with missing value problem
• Our method is interpreted using a probabilistic
  model
• Complexity analysis shows that our method is more
  efficient

17
Epinions Dataset

• 40,163 users who rated 139,529 items with totally
  664,824 ratings
• Rating Density 0.01186
• 18,826 users, representing 46.87 of the
  population, submitted fewer than or equal to 5
  reviews
• The total number of issued trust statements is
  487,183

18
(No Transcript)
19
Metrics

• Mean Absolute Error

20
Comparisons
MAE comparison with other approaches (A smaller
MAE value means a better performance)
PMF CPMF R. Salakhutdinov and A. Mnih (NIPS08)
MMMF J. D. M. Rennie and N. Srebro (ICML05)
21
Impact of Parameters
22
Performance on Different Users

• Group all the users based on the number of
  observed ratings in the training data
• 10 classes 0, 1 - 5, 6 - 10, 11 - 20,
  21 - 40, 41 - 80, 81 - 160, 160 - 320,
  320 - 640, and gt 640,

23
(No Transcript)
24
(No Transcript)
25
Efficiency Analysis

• On a normal PC with Intel Pentium D (3.0 GHz,
  Dual Core) CPU, 1 Giga bytes memory
• When using 99 data as training data
• Less than 20 minutes to train the model
• When using 20 data as training data
• Less than 5 minutes to train the model

26
(No Transcript)
27
Conclusions

• Propose a novel Social Recommendation framework
• Outperforms the other state-of-the-art
  collaborative filtering algorithms
• Scalable to very large datasets
• Show the promising future of social-based
  techniques

28
Future Work

• Kernel representation
• Information diffusion between users
• Distrust information

29
Thanks! Q A Hao Ma Email hma_at_cse.cuhk.edu.hk